Adatrix adaptive trend based optimization in digital advertising
Download as PPT, PDF1 like856 views
Trend analysis is the process of identifying major and minor trends in ad consumption and response behavior. Adaptive trend based optimization analyzes the ad response stream in real time and adapts the ad delivery to respond to identified trends and patterns, helping to optimize ad spend and ROI. Key elements include accurately detecting patterns, predicting patterns early based on past trend data, identifying parameters defining trends, and adapting ad delivery in real time to respond to trends. Tools and techniques from other domains like technical indicators, Fibonacci sequences, and Elliott wave theory from stock markets can inform trend analysis approaches.
Adatrix adaptive trend based optimization in digital advertising
- 1. Adaptive Trend Based Optimization
- 2. What’s the bigger concept ? "With the past, we can see trajectories into the current and future - both catastrophic and creative projections." John Ralston Saul
- 3. 1 DEFINITIONS TREND ANALYSIS & TREND BASED OPTIMIZATION
- 4. Trend Analysis ? General Trend analysis is a structural mapping of expected changes in the behavior of societies, markets, and the consumers who drive them.
- 5. Trend Analysis ? In Our Case Trend analysis is the process which is used to identify major and minor trends in ad consumption and response behavior.
- 6. Trend Based Optimization ? The ad response stream is analyzed in real time and the ad delivery is adapted to respond to the trends (patterns) identified. This will help in optimizing the ad spend and ROI effectively.
- 7. 2 HOW DOES IT HELP ? Value proposition for ADATRIX
- 8. Real Time Targeting How this will help ? Optimizations If in a time window, some trend is detected we can achieve real time optimization by targeting on parameters which define this window. This will provide key trend insights to advertisers which can be used in future campaigns. The trend data will add real-time intelligence to campaigns
- 9. Real Time Bidding How this will help ? Optimizations Trend analysis can also be helpful in determining the bid pricing behavior in the open exchange. Help in knowing the market resistance/support and determining optimal pricings. Can help in price regulations. This can be can an added parameter in our bidding algorithms.
- 10. 3 WHERE DOES IT FIT ? The Flow
- 11. Where does it fit ? Ad is delivered to consumers. Consumers responds by seeing or clicking or filling a form Response stream comes to Adatrix The stream is analyzed in real time for patterns. The scores are updated to reflect the trend This intelligence is used to adapt and serve the next ads
- 12. 4 IMPORTANT QUESTIONS / POINTS Points which need more thinking
- 13. Crucial Elements Pattern detection ( What defines a pattern ?) Pattern prediction ( Based on past trend data the ability to predict a pattern early) Identification of key parameters which define the trend Adapting the ad delivery to respond to the trend in real time
- 14. 5 HOW DO WE DO IT ? Trend Analysis, Algos, Examples and Methodology
- 15. Forex trading Algos for spotting Trends
- 16. 3 Line Break Chart – Forex Trading Example Trend based rules for a three-line break chart: Buy when a white line emerges after three adjacent black lines (a "white turnaround line"). Sell when a black line appears after three adjacent white lines (a "black turnaround line"). Avoid trading in "trendless" markets where the lines alternate between black and white.
- 17. Line Break Chart – Advantage and Disadvantage An advantage of Three Line Break charts is that there is no arbitrary fixed reversal amount. It is the price action which gives the indication of a reversal. The disadvantage of Three Line Break charts is that the signals are generated after the new trend is well under way.
- 18. Technical Indicators – Parameters for Algo
- 19. 6 TREND ANALYSIS - OVERVIEW Trend Analysis Introduction
- 20. Trend Analysis Peaks & Troughs Peak can be defined as the highest high of an uptrend while a Trough can be defined as the lowest low of a downtrend.
- 21. Trend Analysis Defining the Trend Identify Trends within trends Short term trends Consist of either higher highs or lower lows Long term trends Consist of succession of higher peaks or lower troughs
- 22. Trend Analysis The End of a Trend Short Term : Can be defined with 3 bar break Long Term : After a lower peak is made the uptrend ends when a bar breaks through the trough. ( Image above )
- 23. Trend Analysis Trend Line Uptrend: Trend lines for uptrends join the lows Downtrend: Trend lines for downtrends join the highs
- 24. Trend Analysis Trend Channel The picture above shows a downtrend channel Once the line crosses above the channel, signals a possible change in trend
- 25. 7 INSIGHTS FROM STOCK MARKET Fibonacci Sequence and Golden ratios, Elliot Wave Theory
- 26. Use of Fibonacci Sequence Fibonacci sequence can help to forecast support and/or resistance as well as to suggest potential target levels for existing trends & Calculate the timing of future market turning points.
- 27. Use of Fibonacci Sequence Markets display trend and counter trend moves which have relationships with various Fibonacci numbers. The most significant Fibonacci related numbers are 61.8% and 38.2%. Both of these numbers are related to Phi, the golden number. 61.8% = 1/Phi (1.618033988) 38.2% = 1/Phi Squared
- 28. Use of Fibonacci Sequence Moving Average periods are Fibonacci numbers
- 29. Eliott Wave Theory – Market Trends Fibonacci Retracement : corrective waves have been observed to retrace the previous wave by either 38.2%, 50% or 61.8%
- 30. Eliott Wave Theory – Applied to Ad Domain Wave 1 : The Ad campaign starts its first upward phase. Only a relatively small number of people feel interested and engage with the Ad. Wave 2 : Slight correction due to repetitive exposure to same consumer Wave 3 : Increased as more awareness is generated and reach is widened Wave 4 : Slight correction due to over exposure Wave 5 : Maximum awareness generated by the campaign ABC : Beyond this the Ad value and time relevancy decays and the Ad performance drops significantly and then stabilizes to a much lower value
- 31. The Golden Ratio in Trend Analysis Many of the “waves” of movement in the market are related to one another either by the Golden Ratio or by some derivative thereof.
- 32. 8 THE EXTENSION OF CONCEPT How the concept can be applied to provide more value
- 33. Concept Extension Detect variations or waves of sentiment. Detecting a trend is an important part of predicting direction. Is there a shape to changes in response and can it be detected?
- 34. Concept Extension This concept can be applied to multi- channel digital campaigns If a campaign has multiple touch points, the trend data can help in determining how consumers behave in different digital touch points. Cross medium campaign optimization
- 35. 9 WHO ELSE IS DOING ? What already exists ?
- 36. Who else ? It’s a relatively new concept in Ad domain. Already exists extensively in trading markets Its difficult to achieve and validate SurveyAnalytics, The Trend Analysis module allows you to plot Inc aggregated response data over time The dashboard displays real-time data for high-level monitoring of the campaign’s status and for quick MediaMind problem solving. Tracks and report complex user interactions like Dwell Time Automatically optimizes brand campaigns based on AdaptiveAds interactions, engagement, clicks or other key metrics. Tracks media engagement across audiences, providing intelligence for brand messaging.
- 37. 10 TOOLS THAT EXIST ? What already exists ?
- 38. References http://www.jstor.org/pss/1939220 http://www.topgunsoftware.com/resources.html The Undergroundtrader.com Guide to Electronic Trading by Jea Yu http://www.stator-afm.com/stock-trend-analysis.html http://www.adaptiveads.com/product.php http://www.linnsoft.com/tour/threeLineBreakChart.htm http://forex-strategies-revealed.com/develop/three-line- break http://www.esajournals.org/doi/abs/10.2307/1939220
- 39. Thank You !